A jump Markov particle filter for localization of moving terminals in multipath indoor scenarios

This paper describes an efficient Bayesian framework for localization of moving terminals (MT) in wideband wireless networks. In a previous paper, the authors have presented a grid-based technique, based on a hidden Markov model, that used the power delay profiles of the received signals to track the MT position. This grid-based Bayesian method has proved its efficacy in reducing localization errors in realistic indoor environments with multipath effects and mixed line-of-sight/non line-of-sight (LOS/NLOS) conditions. However, the computational power and the memory storage requirements limit its use in practical wireless networks. To improve the computational efficiency, here we propose a jump-Markov particle-filter approach as an extension of the previous work; the LOS/NLOS sight process is the jumping feature that drives the MT motion dynamics, while the particle filter is used to track the MT position. Performance analyses, carried out for realistic multipath indoor environments, show that, with respect to the previous grid-based algorithm, this novel approach greatly reduces the tracking filter complexity still preserving the same localization accuracy. Simulation results prove also the robustness of the proposed method with respect to the uncertainty of sight statistics information.